Holocene displacement field at an emerged oceanic transform-ridge junction: the Husavik-Flatey Fault - Gudfinnugja Fault system, North Iceland

Transcription

1 Holocene displacement field at an emerged oceanic transform-ridge junction: the Husavik-Flatey Fault - Gudfinnugja Fault system, North Iceland Pasquarè Mariotto F. 1, Bonali F.L. 2, 4, Tibaldi A. 2*, Rust D. 3, Oppizzi P. 5, Cavallo A. 2 1 Department of Theoretical and Applied Sciences, Insubria University, Varese, Italy 2 Department of Earth and Environmental Sciences, University of Milan-Bicocca, Milan, Italy 3 Centre for Applied Geoscience, School of Earth and Environmental Sciences, University of Portsmouth, Portsmouth, UK 4 Now at Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano, Milan, Italy 5 Gole della Breggia Geopark, Morbio Inferiore, Switzerland. * Corresponding author: Prof. Alessandro Tibaldi, Department of Earth and Environmental Sciences, University of Milan-Bicocca, P. della Scienza 4, Milan, Italy; alessandro.tibaldi@unimib.it, tel Abstract Our research focuses on Holocene tectonics in a broad area surrounding the junction between the active NW-SE trending Husavik-Flatey transform fault (HFF) and the N-S Gudfinnugja normal fault (GF), an exceptional example of onshore transform-ridge intersection. We mapped 637 minor and major faults, and measured the dip-slip and strike-slip offset components on the major faults. We also mapped 1016 individual fissures, as well as opening directions on the most reliable ones. The results indicate that this portion of the HFF comprises major right-stepping segments, with both normal and right-lateral strike-slip components, linked by local normal faults. The entire GF always shows pure dip-slip normal displacements, with a strong decrease in offset at the junction with the HFF. Fissure opening directions are in the range N45-65 E along the HFF, N90 E along the GF, and N110 E within the area south of the HFF and west of the GF. Fault kinematics and fissure openings suggest a displacement field in good agreement with most of present-day GPS measurements, although our data indicate the possible long-term Holocene effects of the 1

2 superimposition of magma-related stresses on the regional tectonic stresses. The HFF and the GF work together as a structural system able to accommodate differential crustal block motion, and possibly past dyke intrusions Keywords: Gudfinnugja fault; Husavik-Flatey fault; transform fault; normal fault; fissure; Iceland Introduction The active Iceland rift zone runs NNE-SSW across the entire island and, in its northern sector, it is connected with the Kolbeinsey mid-ocean ridge (Fig. 1). This connection is expressed by the parallel right-lateral strike-slip Grimsey, Husavik-Flatey, and Dalvik faults, which make up the 120-km-long Tjörnes transform fault zone (Ward, 1971; Saemundsson, 1974). Most of these faults are offshore, whereas the south-eastern segment of the Husavik-Flatey Fault (HFF) is very well expressed on-land, where the fault has a length of 25 km, strikes NW-SE across the southern part of the Tjörnes peninsula and connects to a set of approximately N-S-striking normal faults and fissures known as the Theystareykir fissure swarm (TFS), part of the North Volcanic Zone of Iceland. Most earthquakes along the rift zone are <M4 (Einarsson and Bjornsson, 1979; Gudmundsson, 1999), whereas transform zones such as the Tjornes are characterised by destructive events with the greatest magnitudes in Iceland (M 6-7), although taking place mostly offshore (Fig. 1) (Tryggvason, 1973; Halldorsson et al., 1996). The on-land HFF shows several indicators of Holocene movement but limited seismicity in comparison to its offshore counterpart (Einarsson, 1991). Four major earthquakes have occurred along the HFF in the past 200 yrs, all offshore. In 1755, an estimated M 7 earthquake took place in Skjalfandi Bay, and in 1838 a M 6.5 event occurred near the westernmost end of the HFF (Metzger et al., 2011). The latter earthquake was followed, in 1872, by two major M 6.5 earthquakes with epicentres close to the towns of Flatey and Husavık (Fig. 1). In particular, Husavik lies directly above the trace of the HFF; Metzger et al. 2

3 (2011) and, assuming that seismic energy has been accumulating since the two large earthquakes of 1872, suggests that the seismic potential of the fault is equivalent to a M w 6.8 ± 0.1 event. Given the high seismic potential of the HFF, detailed field investigations along its trace have been carried out on major faults, dykes, mineral veins (Saemundsson, 1974; Young et al., 1985; Gudmundsson et al., 1993; Fjader et al., 1994; Magnusdottir and Brandsdottir, 2011) and major and minor striated faults (Angelier et al., 2000; Bergerat et al., 1990, 2000; Garcia et al, 2002; Garcia and Dhont, 2005; Bergerat and Angelier, 2008). Previous research suggested that the HFF ends against the Gudfinnugja normal fault (GF in Fig. 2) (Gudmundsson et al., 1993); although recently Hjartardottir et al. (2012) suggested a possible subsurface continuation of the HFF towards the SE on the basis of earthquake epicentres, and fracture distribution at the Krafla fissure swarm. Detailed field work may aid in shedding new light on the intersection of the TFS with the HFF, an exceptional example of a ridge-transform intersection, and a setting that is hardly ever exposed on land. The main purpose of our research has been to assess the tectonic features and long-term (Holocene) kinematics that characterize a broad area centred on the junction between the HFF and the TFS, in order to contribute to understanding the mechanical and kinematic consistency of the these structures, and to evaluate the previously hypothesised prolongation of the HFF towards the SE. Our approach has consisted of mapping all minor and major faults and fissures in the area, augmenting previously published work. Moreover, for the first time we have assessed the dip-slip and strike-slip offset components of the main faults, and determined the opening directions of the more reliable fissures. We have mapped 637 faults and 1016 individual tension fractures, at a metric to kilometric scale. Using these data, we have been able to shed new light on fault geometry and provide, for the first time, fault offset differences in space, fissure geometry and extension directions for the last few thousand years. We have also compared the above with present-day GPS vectors. 3

4 Geologic-tectonic framework The North Volcanic Zone (NVZ), active since 8-9 Ma (Saemundsson, 1974; Young et al., 1985; Bergerat and Angelier, 2008), is made up of five, approximately N-S-striking rift zones, namely, from west to east: the Theystareykir, Krafla, Fremri-Namur, Askja, and Kverkfjöll volcanic systems (Fig. 1) (Gudmundsson, 2000; Hjartadòttir and Einarsson, 2012). Each of these volcanic systems is made up of 5-20 km-wide and km-long fracture swarms and a main volcano (Saemundsson, 1974). The swarms are composed of normal faults, eruptive fissures and tension fractures that strike parallel to the rift direction. Most of the research in this area has focused on the Krafla fissure swarm (e.g., Angelier et al., 1997; Acocella et al., 2000; Dauteuil et al., 2001; Hjartardottir et al., 2012). The TFS is a 10 km-wide stretch of terrain that is dissected by N-Sstriking normal faults, eruptive fissures and marked by the presence of the Theystareykir central lava shield (Opheim and Gudmundsson, 1989; Garcia and Dhont, 2005). Within the TFS, the most prominent structure is the Gudfinnugja Fault (GF), a Holocene normal fault that represents the western edge of the rift system (Fig. 2). Emission of lava flows from the Theystareykir shield was constrained to about 14.5 ka BP in the study area (Slater et al., 2001; Stracke et al., 2003). The latest eruption occurred about 2.4 ka BP, when the Theystareykjahraun lava flows were emplaced, between the central shield and the HFF (Saemundsson et al., 2012). Active vertical deformation at Theystareykir was measured using GPS-based methods (Metzger et al., 2011). The latest major event in the area was the rifting episode that took place within the nearby Krafla volcanic system between 1975 and 1984 (Bjornsson, 1985). During that period, several volcano-tectonic episodes produced consistent displacements marked by horizontal (several meters) and vertical (a few metres) offsets (Tryggvason, 1980, 1984, 1986). The extension was associated with M earthquakes (e.g., Tryggvason, 1980, 1984; Bjornsson, 1985). The HFF strikes 25 km through the Tjörnes peninsula (Fig. 1), as far as the western border of the Holocene rift zone. It is composed of en-échelon, dominantly right-stepping, dextral strike- 4

5 slip fault segments (Gudmundsson 1993, 2007). Near the town of Husavik, the HFF separates Tertiary rocks to the north from Upper Pleistocene rocks to the south (Saemundsson, 1974; Garcia et al., 2002). The fault, first described by Einarsson (1958) and mapped by Saemundsson (1974), displays at the surface vertical offsets reaching 200 m at many sites (Gudmundsson, 1993). Its total vertical displacement may amount to as much as 1400 m (Gudmundsson, 1993), which is in agreement with the 1100 m vertical displacement off the coast of the Flatey fault estimated by Thors (1982). Saemundsson (1974) suggested a right-lateral displacement of as much as 5-10 km, and Young et al. (1985) suggest it may amount to 20 km. Garcia and Dhont (2005) subdivided the HFF into three parallel fault traces and, in agreement with Gudmundsson (1993), identified two major sag ponds along the fault trace, interpreting them as pull-apart basins generated by transtensive displacement. As mentioned above, the HFF connects with N-S-striking normal faults belonging to the TFS (Gudmundsson, 1993); the most spectacular of these interactions is visible where the HFF joins the GF, at an angle of about 60 (Gudmundsson et al., 1993). Near the intersection of the GF and HFF, Gudmundsson et al. (1993, 2007) describe zones of transpression and transtension. The former are marked by fragmented lava blocks forming hills and irregular ridges, while the latter result in pure tensional fractures and small collapse structures. Figure 2 illustrates faults and tension fractures as mapped during previous research efforts in the area, summarized in Saemundsson et al. (2012). Rose diagrams are included, depicting the main strike of faults and fractures: Fault strike ranges from NW to NNE, with a dominant orientation around N-S and NNE; fracture strike ranges from NNW to NNE, with a dominant N-S strike Results In the following sections we describe the data we gathered by field mapping, aided by highresolution satellite imagery. Overall, we recognise 1016 tension fractures, some hundreds of which were also mapped in the field. In measuring their opening directions, we measured only those 5

6 fractures that were distant from fault scarps and that are at least 200 m long, thus avoiding possible local gravity effects on the opening vectors. We selected the most reliable tension fractures (120 in total) and measured in the field their strike and opening direction. We also mapped and measured in the field the fault swarms that make up the GF and HFF (including at their junction), by determining their strike, dip, dip-slip and strike-slip displacement components, kinematics, block rotation and amount of tilting. Although the area has been mapped in the past, our fieldwork enabled us to collect a large amount of new data that can usefully be integrated with the results of previous research efforts. In particular, neither the offsets along the whole length of the GF, nor the opening directions of tension fractures, had previously been assessed in any detail. We describe the area in terms of different kinematics and deformation patterns affecting three sub-areas: 1) the studied section of the HFF, marked by strike-slip deformation structures, 2) the GF, with dip-slip deformation patterns, and 3) the block located south of the HFF and west of the GF (SW-block in Fig. 3). The geological-structural map and rose diagrams of Figure 3 refer to all the structures mapped through both field work and interpretation of high-resolution satellite images, whereas the following figures and graphs show exclusively the structures and data measured directly in the field. As can be observed in the rose diagrams of Figure 3, the overall strike of tension fractures and faults ranges from NW-SE to NNE-SSW, with a principal N-S orientation for fractures and a dominant N-S to NNE-SSW orientation for fault scarps. In more detail, the 784 tension fractures we recognised along the HFF strike in the NNW-SSE to N-S range, with a dominant NNW-SSE strike. The 79 faults strike from NW-SE to NNE-SSW, with a main NW-SE strike and a subsidiary N-S one. Along the rift zone, the 564 tension fractures and 433 faults mostly strike from NNW-SSE to NNE-SSW. However, the dominant strike of the fractures is N-S. In the SW block, the 268 tension fractures and 125 faults strike from N-S to NNE-SSW, with a dominant NNE-SSW orientation. In the next chapters we describe in further detail the data collected in the field in the three sub-areas. 6

7 The Husavik-Flatey Fault The stretch of the HFF that we surveyed in the most detail in the field is made up of three main, WNW-ESE-striking segments (I, III and V in Fig 3). These are made up of single faults linked by two NNW-SSE-striking segments (II and IV in Fig. 3). From Segment II eastwards, the 2.4-ka-old lavas are mostly confined to the downthrown block, although they locally onlap the fault scarps and are involved in the younger deformation events. The 14.5-ka-old lavas crop out on the relatively uplifted block north of the faults. Segment I strikes N123 E and is represented by a rectilinear fault mostly covered by active scree deposits, along which three main landslides of post-lgm age are aligned. At the south-eastern termination of this segment an abrupt change in the structural architecture can be observed: the single, N123 E-striking fault is replaced by a set of faults that strike from N154 E to N180 E (Segment II in Figs. 3 and 4). In its northern part, this fault swarm is 125 m wide and marked by seven, mainly dip-slip fault planes, five of which dip towards the WSW and two in the opposite direction. To the south these faults gradually take on an en-échelon arrangement and eventually form a 20-m wide deformation zone marked by fault scarps facing towards the WSW, with a total dip-slip offset of a few metres. About 110 m west of these faults the footwall block is characterised by the presence of NNW-SSE-striking fissures. In the southernmost portion of the area in Figure 4A the 20-m wide fault zone widens into a fan-shaped fault swarm that extends southwards into a N-Sto NNE-SSW-striking fault zone. East of this fan shaped fault zone, a single surface fault trace marks the HFF that here strikes N114 E (segment III in Fig. 3). It shows evidence of a major right-lateral strike-slip component, plus a normal dip-slip component with the downthrown block to the south. Segment III is 1-km long and disappears in correspondence of another NNW-SSE-striking fault swarm (Segment IV), which is made up of four, approximately 150-m long fault scarps striking from N153 E to N180 E, facing towards the WSW. To the south this fault swarm is abruptly interrupted by Segment V (Fig. 7

8 ), comprising a series of slightly offset fault traces, which will be described from west to east. The first fault strikes N123 E and bears evidence of both right-lateral strike-slip and normal dip-slip components, with the downthrown block to the south. This fault trace is relatively straight up to the point where a left-stepping jump occurs with some overlap (Fig. 5A). Here, another N124 Estriking fault shows clear evidence of right-lateral strike-slip displacement. Further to the SE, a right-stepping geometry, partially overlapping, is clearly observed with a new fault striking N130 E. The area where the left-stepping is observed is marked by domed structures that can be interpreted as push ridges generated by transpressional deformation (contractional overstep), consistent with right-lateral kinematics. By comparison the area where the right-stepping jump occurs is characterised by a depression that can be interpreted as the expression of transtensional deformation (extensional overstep, Fig. 5). Further east, Segment V, although showing a rectilinear surface trace on satellite images and aerial photos, on closer examination in the field is characterised by the presence of systematic en-échelon fractures striking N130 E-140 E. Most of these fractures show pure extension, but some have a right-lateral strike-slip component. Hence, they can be considered as en-échelon, left-stepping fissures and Riedel faults associated with a shear zone. Farther eastwards, Segment V splits into three main fault planes that gradually bend to a N-S strike. Along with the change in fault orientation, the strike-slip component fades away and the dip-slip component takes over as the faults link to the N-S-striking GF. We have also carried out a detailed field survey southeast of the junction across an area of about 2 km, completed by analysis of high-resolution satellite images. Our investigation did not show any evidence of strike-slip faulting at the surface, nor the presence of fault scarps or open fissures parallel to the HFF strike. The only observation here is a slight local re-orientation of a N- S-striking normal fault located about 750 m east of the GF, where a few NW-striking fissures with a left-stepping geometry were also observed

9 The Gudfinnugja Fault (GF) We describe the GF, the westernmost structure in the TFS (Fig. 2), subdividing it into three sections: i) north of the junction with the HFF (1.5 km in length), ii) in close proximity to the junction, and iii) south of the junction (4.5 km long). In the northern section, lavas cropping out are 14.5 ka old on both sides. South of the junction, 14.5-ka-old lavas crop out on the eastern block and 2.4-ka-old lavas occur on the western block. The first evidence of the GF to the north is represented by NNE-striking fissures, south of which the actual GF occurs with an initial offset of about 2 to 3 m (Fig. 6A) along the first 100 m of its length, which increases southwards (e.g. Figs. 6B-C). In the whole northern sector of the GF, lavas are displaced by normal faulting along one single fault scarp, either without any rotation of the footwall block, or with very limited rotation (< 10 ). Assuming an original horizontal topography, as suggested by the surrounding flat area. The deformation style along the GF changes abruptly at the intersection with the HFF, where we observed that the lavas dip about 35 westwards. The dip angle becomes steeper to about 65 a few tens of metres south of the triple junction (Fig. 6D). This monoclinal deformation style, characterised by the hangingwall block dipping in the same direction as the fault plane, was observed along the whole southern sector of the GF (from the triple junction to its southern tip), with dip angles ranging from 40 to 65 (Figs. 6D-F) Vertical offset measurements along the HFF and GF We devoted considerable effort to the quantification (with the highest possible instrumental detail) of offset variations along the HFF and GF. With regard to the former, we were able to assess its vertical offset component manually because fault scarps in the whole studied section are always < 17 m, hence enabling us to record offsets by means of a tape measure. However, along the HFF the quantification of the strike-slip offset component was difficult due to the lack of piercing points. Regarding the GF, as its offsets are much greater than those along the HFF, our methodology 9

10 comprised GPS measurements taken every 100 m as we walked both along the upper surface of the footwall block and the lowermost surface of the hanging-wall block. The GPS altitude data collected were processed to show the variations in offset amounts along the whole length of the GF. Errors are in the order of 5% for tape measurements and 2 m for GPS measurements. In regard to the HFF, within Segment II (Fig. 3), which is made up of a series of parallel faults, the total cumulative fault offset on all the scarps is 22.5 ± m. The parallel faults are marked by pure dip-slip kinematics, hence this value represents the net slip. The dip-slip component of fault offset at Segment III is 5 ± 0.25 m. At Segment IV the parallel fault scarps are from 2 ± 0.1 to 8 ± 0.4 m high each, totalling 20 ± 1 m of net dip-slip offset. The faults of Segment V have a normal dip-slip component from 4 ± 0.2 m to 7 ± 0.35 m. Near the triple junction, where the HFF splits into three main en-échelon faults, each fault has a dip-slip of about 5 ± 0.25 m, and the total dip-slip offset component is in the order of 15 ± 0.75 m. Along the GF, we recorded (Fig. 7) offset values by GPS along a stretch of almost 6 km; from the southern fault terminastion it is possible to notice a sharp increase in offset up to a maximum value of 27 ± 2 m. Northwards, there is a series of irregular variations in offset with a peak at 33 ± 2 m. Most values are between 15 and 22 m. Near the triple junction there is a second peak of 31 ± 2 m. North of the triple junction, there is a local peak of 28 ± 2 m and then dip-slip offset values decrease with some minor variations, mostly in the range m, until the fault fades into a simple fissure with pure dilation Tension fractures Tension fractures have been mapped in the whole study area of Figure 3, but were measured directly in the field only in a zone centred on the triple junction with a radius of about 3 km. For each tension fracture, we assessed: 1) fracture strike, by averaging measurements along each 10-mlong segment; 2) the average opening direction, obtained by measuring at several points the 10

11 azimuth of the line connecting offset piercing points (e.g. in Fig. 8); and 3) the amount of net opening and the along-strike and normal-to-strike components. As shown in Figure 3, the 784 tension fractures in the HFF sub-area strike in the N315 W- N25 E range, with a dominant N315 W strike. In the rift zone sub-area, the 564 tension fractures predominantly strike in a northerly direction. In the SW block, the 268 tension fractures we recognised strike between N0 E and N25 E, with a peak in the N23 E direction. For each of the sub-areas studied we plotted fracture and fault strike vs. longitude (Fig. 9). In Figures 9A and 9B it is possible to observe that all along the HFF fractures strike in the range N E, whereas faults tend to have a smaller strike variation and gradually rotate along longitude. In regard to the other two sub-areas it is worth highlighting the similar distribution between fracture strikes and fault strikes. In the GF rift sub-area (Figs. 9C, 9D), fractures and faults are markedly more frequent east of the longitude that corresponds to the junction between the HFF and the GF. Moreover, fractures and faults in this rift sub-area range from N160 E to N220 E along the GF, whereas they tend to peak at a more northerly strike east of the GF. With regard to the SW-block sub-area we highlight the scarcity or absence of tension fractures and faults within the first 2.5 km west of the GF (Figs. 9E, 9F). Although a few fractures do occur near the GF, the wide fracture field described by Garcia and Dhont (2005) corresponds to a set of lava flow structures belonging to a young lava field that flowed northward (Fig. 2). These structures may be erroneously attributed to post-lava deformation but, in fact, they can be interpreted as accommodation zones that form during lava emplacement under a brittle-plastic regime (for details see Tibaldi, 1996). Regarding fracture opening directions (Fig. 10A and related rose diagrams) along the rift they range from N45 E to N112 E, with a clear predominance of N E-trending directions and a statistical peak at N90 E. Along the HFF, opening directions show two clusters, one around N50-80 E (less represented) and a predominant cluster from N80 E to N100 E. Finally, in the SWblock, opening directions vary from N80 E to N140 E, with a clear predominance at N E and a statistical peak of N112 E. In Figure 10A we also report the trend of GPS spreading vectors, 11

12 as determined by Jouanne et al. (2006), for the years and (trending around N110 E). Also, in order to graphically compare GPS spreading measurements and our field data, we include fracture opening directions and strikes. Figures 10B and C are the result of plotting spreading vector trend and fracture strike vs. latitude and longitude; it is particularly worth noting that opening directions are consistently most frequent in the range N E (Fig. 10B), with an increase towards N E west of the triple junction Volcano-tectonic structures In the studied area we found some structures that can be referred to volcanic and volcanotectonic processes. In the western part there are a few, still recognizable, vents, some of which are associated with small pyroclastic cones occurring on the summit area of the Hofudreidarmuli volcano (Fig. 3). On top of the edifice there is a set of N-S to ENE-WSW-striking fractures (a few tens of metres in length) and normal faults, departing from the cones craters. On account of this observation, and their limited length, these structures can be regarded as due to volcano-tectonic processes such as deformation due to magma upwelling and subsequent magma outpouring with shallow deflation. Less than 0.5 km south of the Hofudreidarmuli volcano there are other vents that, together with those described above, comprise a N-S-trending elongated zone of craters (Saemundsson et al., 2012). In the SW block sub-area, the Theistareykjahraun lava field is also characterised by the presence of several effusion points that appear coeval based on the similar stratigraphic level of the outpoured lavas (Fig. 3). Although more detailed investigations are necessary to assess the complete distribution of vents within this lava field, the vents do appear to be concentrated in a general N-S alignment Discussion Through the collection of new detailed field data we are able to determine the Holocene displacement field in the area surrounding the junction between the HFF and the GF. Our analysis 12

13 of the HFF enabled us to assess that the surface expression of this major transcurrent structure is represented by an en-échelon arrangement of faults rather than a single, continuous, strike-slip fault. We also observed several transtensive and locally transpressive structures, some of which had already been recognized by Gudmundsson (1993). With the purpose of defining how recent tectonics has affected the area, we also quantified in detail for the first time the deformation along the GF and the opening direction of the associated fissures Fault geometry and kinematics The Husavik-Flatey Fault (HFF) We found that the portion of the HFF studied is composed of three main, en-échelon arranged, WNW-ESE-striking segments linked by NNW-SSE-striking fault zones (Fig. 11). The latter are characterised by dip-slip faults with scarps mainly facing towards the WSW. The rightstepping arrangement of the WNW-striking segments, and their right-lateral strike-slip component of displacement, are compatible with local extension manifested as asymmetric pull-apart basins. We interpret the observed geometry in terms of the structural inheritance of older, N-S-striking, dipslip faults that very likely interfered with tectonic movements along the deep-seated HFF wrench fault. In fact, as can be observed in Figure 2, several N-S-striking older faults occur in the Pliocene- Early Quaternary rocks north and south of the studied area. Our interpretation is consistent with previous research elsewhere (Mann, 2007; Mann et al., 2007), which, based on field regional and global case studies of strike-slip faults in complex structural settings, interpreted the formation of restraining and releasing bends as the effect of the interaction between strike-slip fault systems and inherited structures in the basement. In particular, the Jamaican field case discussed in Mann et al. (2007) was interpreted in terms of a left-lateral transcurrent fault interacting with oblique structures, which resulted in the formation of a topographic uplifted area due to local transpression. Based on the model proposed by Mann et al. (2007), we suggest that a major right-lateral structure (as opposed to the Jamaican, left-lateral case) such as the HFF, interacting with older structures, would 13

14 produce the opposite configuration, i.e. the formation of pull-apart basins associated with transtension in a right-stepping geometrical arrangement. The above interpretation is also consistent with recent research by Curren and Bird (2014) that produced a plane-stress finite-strain physical analogue model proving the effect of pre-existing mechanic discontinuities on the evolution of strike-slip fault systems. At a more local scale, along the easternmost portion of the HFF the presence of en-échelon fissures and Riedel faults in a left-stepping arrangement has been observed (e.g. southeastern zone of Figs. 5A and 11). These may be regarded as the earlier structures that develop along a wrench zone; an interpretation consistent with the fact that, although the HFF is relatively old, the build up of younger lava flows hindered the development of a single, well-defined wrench fault plane at the surface. Hence, the observed en-échelon fissures and Riedel faults may be interpreted as an early stage of development of the wrench fault in the uppermost and younger rocks. With regard to the supposed continuation of the HFF east of the triple junction, Hjartardòttir et al. (2012) suggested the buried prolongation of the HFF as far as the Krafla Fissure System (KFS). The above hypothesis was based on two facts: i) earthquake migration from the KFS towards the HFF was observed during the Krafla rifting episode; ii) the central graben of the KFS widens abruptly at the HFF-KFS intersection, and the maximum fracture density in the KFS is accordingly found in this area. Our field survey and interpretation of detailed satellite images up to 2.5 km southeast of the triple junction did not show evidence of strike-slip faulting at the surface and structures parallel to the HFF. This is in agreement with Metzger et al. (2013) who, based on GPS data and back-slip modelling, suggested that the HFF ends at the TFS. Furthermore, Geirsson et al. (2006) and Metzger et al. (2013) pointed out that more than 60% of total relative plate motion of 20.3 mm yr 1 (e.g. DeMets et al., 1994) is accommodated by the offshore Grimsey Oblique Rift, and the HFF dissipates only about 30% of it. Therefore, although we cannot rule out the possible prolongation of the HFF at depth, based on our field data and data from the literature, we suggest that either the HFF ends at the TFS or it extends at depth but has not yet affected the surface. 14

15 The Gudfinnugja normal fault (GF) Our detailed study of the GF, the westernmost structure of the TFS, allowed us to determine that along the 4 km segment south of the triple junction, the fault offset is consistently greater than 20 ± 2 m and peaks at 33 ± 2 m, whereas north of the triple junction, there is a decline in vertical throw with a dominant range of m and a peak of 28 m. Moreover, the northern section of the GF affects 14.5-ka-old lavas on both sides, whereas to the south, 14.5-ka-old lavas crop out on the eastern block and 2.4-ka-old lavas on the western block. This suggests that the actual offset south of triple junction is likely to be at least between 28 and 53 m, taking into account the average 20-m thickness of the 2.4-ka-old lavas. Also, recent GPS data (Jouanne et al., 2006), indicate that in the last few decades the area north of the triple junction has been moving westward at a much lower 378 velocity than the area south of the triple junction. Our results suggest that this differential behaviour occurred over a much longer interval, producing a cumulative vertical offset that is lower north of the triple junction than south of it. We also noted that the deformation style dramatically changes across the triple junction: south of it, a steeply-dipping monocline is associated with the GF along the hanging-wall block, whereas no monocline can be observed north of the triple junction. The occurrence of monoclinal structures flanking normal faults in Iceland has been explained in terms of: i) the friction produced along the fault plane at depth (Gudmundsson et al., 1993); ii) the upward fault growth (Grant and Kattenhorn, 2004), and iii) the original geometrical attitude of lava beds (Sonnette et al., 2010). Our data indicate that south of the triple junction the younger lavas were, at least in part, emplaced against already existing west-facing fault scarps and this implies that the original dip-direction of the lava beds after emplacement was toward the west. However, field evidence also indicates that lavas within the monocline dip as steeply as 65, too steep to correspond to their original attitude. Moreover, locally, monocline lavas display flow direction structures that suggest they have been tilted after emplacement. Since surface fault movements also occurred prior to the emplacement of 15

16 these younger tilted lavas, the model of Grant and Kattenhorn (2004) cannot be applied here because the existence of surface fractures along the GF, earlier than the emplacement of the tilted lavas, is incompatible with the successive upward propagation of a fault plane. Therefore, we suggest that the presence of the monocline here can be interpreted in terms of a combination of the original attitude of the lava beds and post-emplacement deformation Tension fracture opening directions We discuss here our data on tension fracture opening directions as collected in the different sub-areas and attempt a comparison with recent GPS data, although we need to add a note of caution that opening directions derived by tension fracture dilation represent a cumulative offset since 14.5 ka BP or 2.4 ka BP (depending on the affected rock units) and hence can be regarded as representing the long-term displacement field. Starting with the TFS zone, along the whole length of the GF, tension fractures strike mostly N-S (Figs. 3 and 10A) with dominant N90 E-trending opening directions. This geometry does not change north and south of the triple junction, as also shown by Figure 10C. In the sector east of the GF we also observed some northerly-striking fractures, whose geometry does not change across the supposed buried prolongation of the HFF. The Theistareykjahraun lava field is characterised by the presence of several N-S-aligned emission vents that appear to be coeval based on the similar stratigraphic level of the outpoured lavas, suggesting they represent N-S-striking eruptive fissures. This observation, together with the diffuse presence of N-S-striking normal faults and tension fractures along the TFS zone, suggests that this area was fed in the past by N-S-striking dykes, consistent with previous observations by Saemundsson et al. (2012). Observations during episodes of dyke emplacement at volcanic rift zones elsewhere have suggested that there is a relation between dyking and slip increments along existing faults and tension fractures, as well as the opening of new fractures (Abdallah et al., 1979; Bjornsson et al., 1979; Pollard et al., 1983; Wright et al., 2006). This, in turn, provided evidence that faulting and tensional fracturing were triggered 16

17 by the dyke intrusion (Rubin and Pollard, 1988) and not vice-versa. The stress concentration at the advancing dyke tip produces deformation along fault and fracture planes that strike parallel to the dyke plane. Considering the above, we suggest that the long-term deformation along the studied part of the TFS zone may have resulted from the contribution of multiple pre-historic-holocene dyke injections. Since dykes were injected in a N-S direction, the resulting fracture openings here trend E-W. The GPS vectors of Jouanne et al. (2006) regarding the horizontal velocity field of expressed in the Eurasia fixed reference frame (Fig. 12A) reveal a WNW-directed escape of our whole study area. These vectors are oblique with respect to the studied GF zone and may represent a short-term velocity field superimposed on the long-term, E-W-directed deformation that we documented through our fracture opening data. In other words, the dominant direction of deformation obtained by regional GPS and plate tectonic data is WNW-ESE (Fig. 12), but the field data of the structures along the GF recorded a more E-W-trending dilatation, which is probably due to repeated N-S dyke intrusions. In the SW-block (south of the HFF and west of the GF) we documented a 2.5-km-wide N-S strip, immediately west of the GF, characterised by the absence of tension fractures. Since this area is adjacent to the active rift zone subjected to magmatic intrusions, we can interpret the observed absence of extensional structures as the possible consequence of local, horizontal compressive stresses induced by dyke intrusions along the rift. Previous studies elsewhere in fact (Gudmundsson, 2003; Gudmundsson and Loetveit 2005; Maccaferri et al., 2013), suggested that the magmatic overpressure of intruding dykes along rift zones induces a local stress field perturbation with lateral horizontal compressive stress. Farther west in the SW-block (at a distance > 2.5 km from the GF) tension fractures do occur, but here they have a dominant N0-10 E strike and N110 E-trending opening direction (Fig. 10) that indicate a small component of right-lateral oblique opening. If we compare the distribution of fracture strikes directly measured in the field in the SW-block (Fig. 10) with fracture strikes shown for the same SW-block in Figure 3, we observe that the latter strike more NNE-SSW. This is due to 17

18 the fact that the rose diagrams of Figure 3 also include the structures mapped by means of satellite images in the southwesternmost part of the area, where most tension fractures and faults strike NNE-SSW. The long-term opening direction in the SW-block is identical to the average direction derived from short-term GPS data provided by Jouanne et al. (2006) for the years (Fig. 12B). Metzger et al. (2013), studying the present kinematics of the Tjörnes Fracture Zone, suggested that the overall spreading direction ranges from N109.4 E to N115.1 E, which is similar to our average N110 E-trending opening direction. Our long-term opening directions are also consistent with horizontal GPS velocities measured all over Iceland outside the rift zone (Geirsson et al., 2006; Árnadóttir et al., 2009), and with plate motion models such as REVEL (Sella et al., 2002) (Fig. 12D). This may be explained in terms of horizontal, magma-induced compressive stress fading away at distance from the rift zone, where the regional tectonic stress takes over in controlling the opening direction of fractures. Along the HFF, tension fractures strike between N315 W and N-S, mainly N337 W (consistent with Gudmundsson, 1993), and their opening directions are mostly N E. In the sector north of the HFF we observed very few northerly-striking fractures. Jouanne et al. (2006) show that the modulus of velocity vector is greater south of the HFF and west of the GF in the fixed European frame. The GPS horizontal data illustrated in Metzger et al. (2013) indicate (Fig. 11C) that the tectonic block south of the HFF and west of the GF is stable with respect to a fixed North America frame, whereas the block north of the HFF moves towards the ESE. By merging all the data it is possible to illustrate the long-term Holocene displacement field. In Figure 13 all the fracture opening directions have been averaged and plotted as extension directions (dashed white line). The areas of larger measurement frequency have been represented by means of diverging blue arrows, which indicate the local dominant extension direction. The resulting displacement field is characterised by quite homogeneous E-W orientations to the east in block 1, which coincides with the main rift zone. Westwards, the displacement orientations bend to N110 E directions. In order to account for this rotation it is necessary to insert another structural 18

19 boundary along which a change in the dominant extension direction might occur, separating the SW-block into sub-blocks 3 and 4. This boundary may coincide with the approximately 1-km-wide swarm of Holocene faults and fractures located west of the Theistareykjahraun lava field (Figs. 2 and 3). These displacement directions are coherent with the long-term and present-day "escape" of the southwestern portion of the study area bounded to the north by a right-lateral transtensional shear zone, coinciding with the HFF. It is also consistent with the presence of a right-lateral strikeslip component, and a normal dip-slip component, along the N E strands of Segments I, III and V of the HFF, as well as with the dominant normal dip-slip motions along Segments II and IV Conclusions With the purpose of assessing how recent and active tectonics have been working in the area of the junction between the HFF and the TFS, based on our highly-detailed field survey we are able to point out the following: - Although most seismicity at the HFF is located in the offshore segment, there is also evidence of recent motion along the on-land segment based on the presence of fresh morphologies such as fault scarps and fractures, as well as evidence of offsets in Holocene lavas. Holocene motion of the HFF in the study area is marked by a right-lateral strike-slip component that is difficult to quantify in the field, and a dip-slip component that gradually decreases towards the GF. - The studied portion of the HFF is composed of three main, right-stepping WNW-ESEstriking, fault segments. The areas between these segments are subject to extension associated with the presence of releasing bends within NNW-SSE-striking fault zones. Thewse bends are characterised by dip-slip normal faults, mainly (90%) dipping to the WSW. We interpret, in agreement with previous research (Mann, 2007; Mann et al., 2007; Curren and Bird, 2014) the observed geometry in terms of the structural inheritance of older, N-S-striking dip-slip faults that very likely interfered with tectonic motion along a deep- 19

20 seated wrench fault. At one location we found a major left-stepping contractional overstep, whereas several small-scale left-stepping Riedel shears are present in the easternmost part of the HFF. - With regard to the possible continuation of the HFF east of the junction with the GF, we cannot rule out its prolongation at depth but, based on our own field data and data from the literature, we suggest that either the HFF ends at the GF or it prolongs farther east but has not yet affected the surface. - Our study of the GF along its whole length shows that, south of the triple junction, fault offsets are much greater than north of it, being also in part obscured by the onlap of late Holocene lava flows (total estimation of m south vs m north). Additionally, the deformation style along the fault dramatically changes from monocline tilting along the whole length south of the triple junction, to a single fault scarp with no, or very limited, rotation of the hanging-wall block to the north. The geometry and kinematics of the normal faults and tension fractures along the GF indicate long-term E-W-trending dilation compatible with the emplacement of N-S-striking dykes over the Holocene. - The absence of recent extensional features in a 2.5-km-wide strip west of the GF and south of the triple junction can be interpreted as the possible consequence of local, horizontal compressive stress induced by dyke intrusions along the rift. - Tension fractures and normal faults do occur farther west of the 2.5-km-wide strip, which might be explained in terms of horizontal, magma-induced compressive stress that fades away at distance from the rift zone. At this distance regional tectonic stress takes over and affects the opening of tension fractures compatibly with present-day motions of the area along a N110 E direction, consistent with recent GPS measurements. - All these data suggest that the HFF and the GF worked during the Holocene as a structural system able to accommodate differential crustal block motions and possible past dyke 522 intrusions. 20

21 Acknowledgments We are grateful to Joao Hippertt (Editor) and Agust Gudmundsson for their precious comments and suggestions. We also thank the Iceland Meteorological Office the seismic data of the studied area. This work is a contribution to the International Lithosphere Program - Task Force II. We also acknowledge Andrea Rovida for his helpful insights into the historical seismology of the area. References Abdallah, A., Courtillot, V., Kasser, M., LeDain, A.Y., Lepine, J.C., Robineau, B., Ruegg, J.C., Tapponnier, P., Tarantola, A., Relevance of Afar seismicity and volcanism to the mechanics of accreting plate boundaries. Nature, 282, Acocella, V., Gudmundsson, A., Funiciello, R., Interaction and linkage of extension fractures and normal faults: examples from the rift zone of Iceland. J. Struct. Geol. 22, Angelier, J., Bergerat, F., Dauteuil, O., Villemin, T., Effective tension-shear relationship in extensional fissure swarm, axial rift zone of northeastern Iceland. J. Struct. Geol. 19, Angelier, J., Bergerat, F., Homberg, C., Variable coupling explains complex tectonic regimes near oceanic transform fault: Flateyjarskagi, Iceland. Terra Nova 12, Árnadóttir, T., Lund, B., Jiang, W., Geirsson, H., Björnsson, H., Einarsson, P., Sigurdsson, T., Glacial rebound and plate spreading: results from the first countrywide GPS observations in Iceland. Geophys. J. Int. 177, Bergerat F., Angelier J., Immature and mature transform zones near a hot spot: The South Iceland Seismic Zone and the Tjörnes Fracture Zone (Iceland). Tectonophysics 447, Bergerat, F., Angelier, J., Villemin, T., Fault systems and stress patterns on emerged oceanic ridges: a case study in Iceland. Tectonophysics 179, Bergerat, F., Angelier, J., Homberg, C., Tectonic analysis of the Husavık-Flatey Fault (northern Iceland) and mechanisms of an oceanic transform zone, the Tjornes Fracture Zone. Tectonics 19, Bjornsson, A., Dynamics of crustal rifting in NE Iceland. J. Geophys. Res. 90, 10,151 10,162. Bjornsson, A., Johnsen, G., Sigurdsson, S., Thorbergsson, G., Tryggvason, E., Rifting of the plate boundary in north Iceland J. Geophy. Res. 84, Curren, I.S., Bird., P., Formation and suppression of strike-slip fault systems, Pure Appl. Geophys. 171,

22 Dauteuil, O., Angelier, J., Bergerat, F., Verrier, S., Villemin, T., Deformation partitioning inside a fissure swarm of the northern Icelandic rift. J. Struct. Geol. 23, DeMets, C., Gordon, R.G., Argus, D.F. Stein, S., Effect of recent revisions to the geomagnetic reversal time scale on estimates of current plate motions, Geophys. Res. Lett. 21, Einarsson, T., A survey of the geology of the area Tjiirnes-Bardardalur in northern Iceland, including paleomagnetic studies. Sot. Sci. Iceland, Rit 32, Einarsson, P., Earthquakes and present-day tectonism in Iceland, Tectonophysics 189, Einarsson, P., Bjornsson, S., Earthquakes in Iceland. Jökull 29, Fjader, K., Gudmundsson, A., Forslund, T., Dikes, minor faults and mineral veins associated with a transform fault in North Iceland. J. Struct. Geol. 16, Garcia, S., Dhont, D., Structural analysis of the Húsavík Flatey Transform Fault and its relationships with the rift system in Northern Iceland. Geodin. Acta 18, Garcia, S., Angelier, J., Bergerat, F., Homberg, C., Tectonic analysis of an oceanic transform fault zone revealed by fault slip data and earthquake focal mechanisms: the Husavik Flatey Fault, Iceland. Tectonophysics 344, Geirsson, H., Árnadóttir, T., Völksen, C., Jiang, W., Sturkell, E., Villemin, T., Einarsson, P., Sigmundsson, F., Stefánsson, R., Current plate movements across the Mid-Atlantic Ridge determined from 5 years of continuous GPS measurements in Iceland. J. Geophys. Res. 111, B09407, doi: /2005jb Grant, J. V., Kattenhorn, S. A., Evolution of vertical faults at an extensional plate boundary, southwest Iceland. J. Struct. Geol. 26, Grünthal, G., Wahlström, R., The European Mediterranean Earthquake Catalogue (EMEC) for the last millennium. J. Seismol. 16, Gudmundsson, A., Tectonics of the Thingvellir fissure swarm, SW Iceland. J. Struct. Geol. 9, Gudmundsson, A., Fluid overpressure and stress drop in fault zones, Geophys. Res. Lett. 26, Gudmundsson, A., Dynamics of volcanic systems in Iceland: Example of tectonism and volcanism at juxtaposed hot spot and mid-ocean ridge systems. Ann. Rev. Earth Planet. Sci. 28, Gudmundsson, A., Surface stresses associated with arrested dykes in rift zones. Bull. Volcanol. 65,

23 Gundmundsson, Infrastructure and evolution of ocean-ridge discontinuities in Iceland. J. Geodyn. 43, Gudmundsson, A., Loetveit, I. F., Dyke emplacement in a layered and faulted rift zone. J. Volcanol. Geoth. Res. 144, Gudmundsson, A., Brynjólfsson, S., Jónsson, M.Th., Structural analysis of a transform faultrift zone junction in North Iceland. Tectonophysics 220, Halldorsson P.T., Skaftadottir T., Gudmundsson G., A new catalogue of earthquakes in Iceland, In ESC XXV general Assembly, Reykjavik, Iceland. Hjartardóttir, Á. R., Einarsson, P., The Kverkfjöll fissure swarm and the eastern boundary of the Northern Volcanic Rift Zone, Iceland. Bull. Volcanol. 74, Hjartardóttir, Á. R., Einarsson, P., Bramham, E., Wright, T. J., The Krafla fissure swarm, Iceland, and its formation by rifting events. Bull. Volcanol. 74, Jouanne, F., Villemin, T., Berger, A. Henroit, O Rift-transform junction in North Iceland: rigid blocks and narrow accommodation zones revealed by GPS Geophys. J. Int. 167, Maccaferri, F., Rivalta, E., Passarelli, L., Jónsson, S., The stress shadow induced by the Krafla rifting episode. J. Geophys. Res.: Solid Earth, 118, Magnúsdóttir, S., Brandsdóttir, B., Tectonics of the Þeistareykir fissure swarm. Jökull 61, Mann, P., Global catalogue, classification and tectonic origins of restraining- and releasing bends on active and ancient strike-slip fault systems. In: Cunningham, W. D. & Mann, P. (eds). Geological Society of London, Special Publication 290, Mann, P., C. Demets, C., Wiggins-Grandison M., Toward a better understanding of the Late Neogene strike-slip restraining bend in Jamaica: geodetic, geological, and seismic constraints. In: Cunningham, W. D. & Mann, P. (eds). Tectonics of Strike-Slip Restraining and Releasing Bends. Geological Society, London, Special Publications 290, Metzger, S., Jonsson, S., Geirsson, H., Locking depth and slip-rate of the Husavık Flatey fault, North Iceland, derived from continuous GPS data Geophys. J. Int. 187, Metzger, S., Jonsson, S., Danielsen, G., Hreinsdottir, S., Jouanne, F., Giardini, D., Villemin, T., Present kinematics of the Tjörnes Fracture Zone, North Iceland, from campaign and continuous GPS measurements. Geophys. J. Int. 192,

24 Opheim, J. O., Gudmundsson, A., Formation and geometry of fractures, and related volcanism, of the Krafla fissure swarm, northeast Iceland. Geol. Soc. Am. Bull. 101, Pollard, D.D., Delaney, P.T., Duffield, W.A., Endo, E.T., Okamura, A.T., Surface deformation in volcanic rift zones: Tectonophysics 94, Rögnvaldsson, S.T., Gudmundsson, A., Slunga, R., Seismotectonic analysis of the Tjörnes Fracture Zone, an active transform fault in north Iceland. J. Geophys. Res. 103(B12), Rubin, A. M., Pollard, D. D., Dike-induced faulting in rift zones of Iceland and Afar. Geology 16, Saemundsson, K., Evolution of the axial rifting zone in northern Iceland and the Tjornes fracture zone. Geol. Soc. Am. Bull. 85, Saemundsson, K., Hjartarson, A., Kaldal, I., Sigurgeirsson, M.A., Kristinsson, S.G., Vikingsson, S., Geological map of the Northern Volcanic Zone, Iceland. Northern Part 1: Reykjavik: Iceland GeoSurvey and Landsvirkjun. Sella, G. F., T. H. Dixon, A. Mao, REVEL: A model for Recent plate velocities from space geodesy, J. Geophys. Res. 107(B4), 2081, doi: /2000jb Slater, L., McKenzie, D., Grönvold,K., Shimizu, N., Melt generation and movement under Theistareykir, NE Iceland. J. Petrol. 42, Sonnette, L., Angelier, J., Villemin, T., Bergerat, F., Faulting and fissuring in active oceanic rift: Surface expression, distribution and tectonic volcanic interaction in the Thingvellir Fissure Swarm, Iceland. J. Struct. Geol. 32, Stefansson, R., Gudmundsson, G.B., Halldorsson, P., Tjörnes fracture zone. New and old seismic evidences for the link between the North Iceland rift zone and the Mid-Atlantic ridge. Tectonophysics 447, Stracke, A., Zindler, A., Salters, V. J. M., McKenzie, D., Blichert-Toft, J., Albarède, F., Grönvold, K., Theistareykir revisited. Geochem. Geophys. Geosyst. 4, 8507, doi: /2001gc Stucchi M., Rovida A., Gomez Capera A.A, Alexandre P., Camelbeeck T., Demircioglu M.B., Gasperini P., Kouskouna V., Musson R.M.W., Radulian M., Sesetyan K., Vilanova S., Baumont D., Bungum H., Fäh D., Lenhardt W., Makropoulos K., Martinez Solares, J. M., Scotti O., Živcić M., Albini P., Batllo J., Papaioannou C., Tatevossian R., Locati M., Meletti C., Viganò D., Giardini D., The SHARE European Earthquake Catalogue (SHEEC) J. Seismol. 17,

25 Thors, K., Shallow seismic stratigraphy and structure of the southernmost part of the Tjornes Fracture Zone. Jokull 32, Tibaldi A., Non-tectonic faulting: examples from late Quaternary trachytes of Ischia Island and basalts of Mt. Etna, Italy. Acta Vulcanologica 8, Tryggvason, E., Seismicity, earthquake swarms and plate boundaries in the Icelandic region. Seismol. Soc. Am. Bull. 63, Tryggvason, E., Subsidence events in the Krafla area, North-Iceland, Journal of Geophysics-Zeitschrift fur Geophysik 47(1-3), Tryggvason, E., Widening of the Krafla fissure swarm during the volcanotectonic episode. Bulletin Volcanologique 47, Tryggvason, E., Multiple magma reservoirs in a rift zone volcano: Ground deformation and magma transport during the September 1984 eruption of Krafla, Iceland. J. Volcanol. Geoth. Rese. 28, Young, K.D., Jancin, B., Orkan, N.I., Transform deformation of tertiary rocks along the Tjornes Fracture Zone, North Central Iceland. J. Geophys. Res. 90, Ward, P.L., New interpretation of the geology of Iceland. Geol. Soc. Am. Bull. 82, Wright, T.J., Ebinger, C., Biggs, J., Ayele, A., Yirgu, G., Keir, D., Stork, A., 2006, Magmamaintained rift segmentation at continental rupture in the 2005 Afar dyking episode. Nature 442,

26 Figure captions Figure 1. Tectonic setting of northeastern Iceland. The mid-atlantic Ridge is here offset by the Husavık-Flatey Fault and the Grimsey Lineament (after Garcia and Dhont, 2005 and Metzger et al., 2013). Yellow stars indicate location of Mw>6 historical earthquakes with indication of magnitude and year (after Stefansson et al., 2008; Metzger et al., 2011; Grünthal and Wahlström, 2012; Stucchi et al, 2012). Orange stripes represent volcano-tectonic rift zones with their names and main surface fault traces (after Rögnvaldsson et al., 1998). Black triangles represent the main Quaternary central volcanoes. Inset shows location of the area and the main volcanic zones of Iceland. Box locates Figure 2. DAL = Dalvik Lineament; FL = Town of Flatey; GRL = Grimsey Lineament; HFF = Husavik-Flatey Fault; KFS = Krafla Fissure Swarm; KR = Kolbeinsey Ridge (segmented line); KR = Krafla volcano (triangle); SB = Skjalfandi Bay; TH = Theistareykir shield volcano; TJ = Tjornes Fracture Zone; HU = Town of Husavik. Figure 2. Fault scarps and tension fractures as mapped during previous studies in the area, along with their rose diagrams. See text for data description. Location in figure 1. HFF = Husavik-Flatey Fault; GF = Gudfinnugja Fault; Hö = Höskuldsvatn pull-apart; Bo = Botnsvatn pull-apart. Data of earthquakes (1994-today) have been provided by Icelandic Meteorological office ( Inset shows location of Figure 3. Figure 3. Detailed geological map of the study area (redrawn after Saemundsson et al. 2012) with the faults and tension fractures identified in the present paper by means of both field work and satellite image interpretation. Rose diagrams are provided for all the identified 637 fault scarps and 1016 tension fractures. Segments I to V indicate the zones of the Husavik-Flatey fault discussed in the text. The dashed black lines separate the three sub-areas studied. Upper box locates Figure 4 of Segment II, and lower box locates Figure 5 of part of Segment V. Figure 4. A. Detailed map of the normal fault zone of Segment II of the Husavik-Flatey Fault based on field data. B. and C. Photo looking SE and interpretation of the en-échelon pattern of normal faults. D. General view of the westernmost NNW-SSE-striking normal fault. Location shown in Figure 3. Figure 5. A. Detailed map of part of Segment V near the eastern termination of the Husavik-Flatey Fault based on field data. B. Examples of pressure ridges. D. Example of right-lateral strike-slip 26

27 motions. E. ESE-striking fault segment showing both transtensional and transpressional features in correspondence of slight changes in fault strike, compatible with right-lateral strike-slip kinematics. Location shown in Figure 3. Figure 6. Examples of different styles of deformation along the Gudfinnugja Fault. Left column shows sites located north of the junction with the HFF, right column south of it. See text for details. Figure 7. Graph showing fault scarp offsets along the Gudfinnugja Fault. GPS measurements have been taken every 100 m. Error is in the order of ± 2 m. Figure 8. Examples of fractures measured in the studied area: A. Pure extensional offset; B. Opening with a subordinate strike-slip component; C. Dominant strike-slip component; D. Compressional fracture typical of transpressional zones. Figure 9. Graphs of tension fracture and fault scarp strike vs. Longitude in each of the three subareas of Figure 3. Here we plotted only the structures measured directly in the field, the structures seen on satellite images have not been included. See text for details. Figure 10. A. Map of the study area with tension fracture opening directions, main faults, GPS spreading directions (from Jouanne et al., 2006) and average fracture strikes. Rose diagrams show fracture strikes (black) and opening directions (blue) in the three sub-areas. B. Graph showing tension fracture opening directions (blue circles) and spreading vector directions (yellow diamonds) vs. longitude. C. Graph showing tension fracture opening directions (blue circles) and spreading vector directions (yellow diamonds) vs. latitude. Figure 11. Sketch map of the main structures identified along the Husavik-Flatey Fault, with an interpretation of their overall kinematics. Figure 12. A. 1997/1999 and B. 1999/2002 GPS horizontal velocity fields in NE Iceland calculated in an Eurasia fixed reference frame. Ellipses correspond to a 2σ level (95 % confidence) (modified after Jouanne et al., 2006). C. GPS horizontal velocities in a fixed North American frame (95 % confidence). Green lines outline fracture swarms and central volcanoes; blue lines indicate the segments of the interseismic deformation model (modified after Metzger et al., 2013). D. GPS horizontal velocity vectors (black arrows) with respect to a local fixed reference frame. The white 27

28 arrows show the predicted REVEL plate motions. Ellipses represent the 2σ confidence level. Inset shows an enlarged map of the active deformation zone in SW Iceland (modified after Geirsson et al., 2006). Figure 13. Displacement field reconstructed by averaging the opening directions (dashed lines) of the tension fractures measured in the field and taking into account the kinematics and distribution of the faults. The boundaries of the various blocks (numbered 1 to 4) have been determined based on the location, kinematics and geometry of the main faults. Diverging blue arrows provide average opening direction where there is the maximum density of field measurements. Black arrows provide the vector of rigid body translation applied to the various blocks as derived by the average displacements measured in the field. 28

29 761 29

30 762 30

31 763 31

32 764 32

33 765 33

34 766 34

35 767 35

36 768 36

37 769 37

38 770 38

39 771 39

40 772 40

41 773 41